/fakereturns

An R package to generate fake stock portfolio data sets.

Primary LanguageRGNU General Public License v3.0GPL-3.0

fakereturns

A toy R package to generate fake but realistic data for various analytic purposes.

Very Alpha, not ready for use.

Overview

The idea is to

  1. Start with a a set of tickers, a begin, and end date that defines a period of time.
  2. Get a number of random dates from this period of time, and
  3. Assign the tickers to these dates.
  4. Retrieve historical stock ticker information for (ticker, date) tuples
  5. Arbitrarily assign BOUGHT or SOLD status

Now we have a dataset to analyze.

Examples

library(fakereturns)

d <- fake.some.portfolio.data(
     start.date= '2015-01-01', 
	 end.date=Sys.Date(), 
	 n = 40,
	 tickers = c("AAPL", "NFLX", "GOOG"),
	 pct.bought = 0.5,
	 purchase.ceiling = 5000
	 
	 )

Installation

if (!require("devtools")) {
  install.packages("devtools")
}
devtools::install_github("aaelony/fakereturns", dependencies = TRUE, build_vignettes = FALSE)

TODO

  • offer ways (e.g. FIFO, LIFO) to match BOUGHT with SOLD
  • compute IRR
  • freq distribution of weekdays

Changelog

Version 0.0.5

Breaking changes, now returns one data.table.

  • compute shares sold with dates occuring after the purchase

Version 0.0.4

adding weekday to generated dates